Endothelial Monocyte-Activating Polypeptide-II Induces BNIP3-Mediated Mitophagy to Enhance Temozolomide Cytotoxicity of Glioma Stem Cells via Down-Regulating MiR-24-3p

Preliminary studies have shown that endothelial-monocyte-activating polypeptide-II (EMAP-II) and temozolomide (TMZ) alone can exert cytotoxic effects on glioma cells. This study explored whether EMAP-II can enhance the cytotoxic effects of TMZ on glioma stem cells (GSCs) and the possible mechanisms associated with Bcl-2/adenovirus E1B 19 kDa protein-interacting protein 3 (BNIP3)-mediated mitophagy facilitated by miR-24-3p regulation. The combination of TMZ and EMAP-II significantly inhibited GSCs viability, migration, and invasion, resulting in upregulation of the autophagy biomarker microtubule-associated protein one light chain 3 (LC3)-II/I but down-regulation of the proteins P62, TOMM 20 and CYPD, changes indicative of the occurrence of mitophagy. BNIP3 expression increased significantly in GSCs after treatment with the combination of TMZ and EMAP-II. BNIP3 overexpression strengthened the cytotoxic effects of EMAP-II and TMZ by inducing mitophagy. The combination of EMAP-II and TMZ decreased the expression of miR-24-3p, whose target gene was BNIP3. MiR-24-3p inhibited mitophagy and promoted proliferation, migration and invasion by down-regulating BNIP3 in GSCs. Furthermore, nude mice subjected to miR-24-3p silencing combined with EMAP-II and TMZ treatment displayed the smallest tumors and the longest survival rate. According to the above results, we concluded that EMAP-II enhanced the cytotoxic effects of TMZ on GSCs' proliferation, migration and invasion both in vitro and in vivo.


INTRODUCTION
Glioma is the most common primary invasive tumor of the central nervous system (Ostrom et al., 2015;Louis et al., 2016). The National Comprehensive Cancer Network (NCCN) Clinical Practice Guidelines state that the current standard of care with respect to glioma management includes surgical resection followed by radiotherapy with concomitant and adjuvant chemotherapy with temozolomide (TMZ) (Stupp et al., 2005;Nabors, 2016). The anti-cancer effects of TMZ, an oral second-generation alkylating agent, are facilitated by its rapid hydrolysis into 5-(3-methyltriazen-1yl) imidazole-4-carboxamide (MITC) under physiological pH conditions (Pace et al., 2003). TMZ has been shown to confer more survival benefits than other chemotherapy drugs. However, the European Organization for Research and Treatment of Cancer/National Cancer Information Center (EORTC/NCIC) trial shows that the median survival time of patients treated with TMZ was 19 months, while the median survival time of patients treated without TMZ was 14.6 months (Rapp et al., 2015). Glioma stem cells (GSCs) are individual glioma founder cells with the capacity for infinite self-renewal and long-term proliferation, as well as the capacity for neurosphere formation and multipotential differentiation (Bao et al., 2006). It has been shown that GSCs is more resistant to conventional therapies than its differentiated counterpart. It has also proved that functional and some mechanistic evidence of the conversion of non-GSCs into GSCs both in vitro and in vivo after exposure to TMZ. Therefore One of the most significant impediments to the success of glioma chemotherapy is the presence of GSCs (Auffinger et al., 2014).
Endothelial-monocyte-activating polypeptide-II (EMAP-II) is a 22-kDa secretory polypeptide protein that was originally isolated from murine methylcholanthrene A-induced fibrosarcoma. It inhibits tumor angiogenesis and causes tumor cell apoptosis without effecting normal cells and thus has been applied in cancer therapy (Berger et al., 2000a,b). EMAP-IIinduced autophagy has been shown to kill glioma cells and GSCs (Ma et al., 2015;Chen et al., 2016). Exogenous EMAP-II could enhance the synergic effects of bortezomib with gemcitabine by cleaving caspase-3 in pancreatic cancer cells (Awasthi et al., 2013). However, whether EMAP-II enhances the cytotoxic effects of TMZ on GSCs remains unknown.
The process of mitochondrial turnover predominantly comprises autophagic sequestration and the delivery of mitochondria to lysosomes for hydrolytic degradation, a phenomenon known as mitophagy (Tolkovsky et al., 2002;Devin and Rigoulet, 2007). Mitophagy is a process in which damaged mitochondria are degraded to maintain cell metabolic homeostasis. However, mitophagy may also entail the clearance of healthy mitochondria in response to hypoxia and starvation, causing cellular metabolic dysfunction and even cell death Kurihara et al., 2012). Bcl-2/adenovirus E1B 19 kDa protein-interacting protein 3 (BNIP3) is a member of the Bcl-2 subfamily of death-inducing proteins and has been found to interact with adenoviruses E1B-19K (Bellot et al., 2009). BNIP3 affects mitophagy by interacting with mitochondria with BNIP3 transmembrane Abbreviations: EMAP-II, endothelial-monocyte-activating polypeptide-II; TMZ, temozolomide; GSCs, glioma stem cells; EORTC/NCIC, European Organization for Research and Treatment of Cancer/National Cancer Information Center; MITC, 5-(3-methyltriazen-1yl) imidazole-4-carboxamide; BNIP3, Bcl-2/adenovirus E1B 19 kDa protein-interacting protein 3; ATG5, autophagy-related protein 5; ATG7, autophagy-related protein 7; EGF, epidermal growth factor; bFGF, basic fibroblast growth factor; FBS, fetal bovine serum; IDVs, integrated density values; NCCN, National Comprehensive Cancer Network. domains (Zhang et al., 2008). It has been reported that TMZ can induce mitophagy in the U87 cell line (Lin et al., 2012), but the specific mechanism underlying this effect has not yet been elucidated. Whether TMZ exerts cytotoxic effect on GSCs though BNIP3-mediated mitophagy, as well as the mechanism though which TMZ exerts its effects, remain unclear.
MicroRNAs (miRNAs) are a family of endogenous, small (18-25 nucleotides) non-coding single-stranded RNA molecules found in all eukaryotic cells. Recent studies have shown that miRNAs can indirectly adjust autophagy, as well as mitophagy . MiR-27a and miR-27b regulate autophagic clearance of damaged mitochondria by targeting PINK1 (Kim et al., 2016). MiR-24-3p has been found to be significantly overexpressed in oral squamous cell carcinoma (Lin et al., 2010) and promotes glioma cell proliferation (Xu et al., 2013). We found that EMAP-II induced upregulation of autophagy-related protein 5 (ATG5) and autophagy-related protein 7 (ATG7) to promote autophagy that killed glioma cells by down-regulating miR-20a(12). However, whether EMAP-II induces mitophagy in GSCs via miR-24-3p and enhances the cytotoxic effects of TMZ on GSCs has not been reported.
This study aims to verify whether low-dose EMAP-II enhances the cytotoxic effects of TMZ on GSCs and explores the mechanism by which EMAP-II induces BNIP3-dependent mitophagy via miR-24-3p down-regulation to enhance the inhibitory effects of TMZ on GSC malignant biological behavior.

Cell Lines and Cell Culture
The human malignant glioma cell lines U-87 and U-251 and the human embryonic kidney cell line HEK293T were purchased from the Shanghai Institutes for Biological Sciences Cell Resource Center. All the cells were cultivated in DMEM supplemented with 10% fetal bovine serum (FBS, Life Technologies Corporation, Paisley, UK), and all the cells were incubated in a moist incubator at 37 • C with 5% CO 2 and refreshed medium every 48 h.
For differentiation experiments, spheres were dissociated by for a single cell suspension and seeded (2.5 × 104 cells/cm2) onto glass coverslips coated with poly-L-ornithine (BD Biosciences, Franklin Lakes, NJ, USA) in the medium without bFGF and EGF but containing 10 % FBS and were incubated for 7-11 days.

Cell Viability Assay
Cell Counting Kit-8 (CCK8, Dojindo, Tokyo, Japan) was used for cell cytotoxicity and proliferation assay. Cells were seeded in 96well plates at a density of 5,000 cells per well, and 10 µl of CCK8 was added into each well at 24, 48, 72 h after DMSO or TMZ treatment (0, 100, 200, 400, 600, 800, 1000, 1200 µM). As the cells incubated at 37 • C for 4 h, the absorbance was measured at a wave length of 450 nm every 30 min.

Cell Migration and Invasion Assay
Cell migration and invasion was measured using a transwell permeable chamber (6.5 mm in diameter, 8 µm pore size, Corning Incorporate, Corning, NY, USA). Cells were suspended in 100 µl of serum-free medium and then added to the upper chamber (without Matrigel for cell migration assay) or seeded in a Matrigel-pre-coated upper chamber (for cell invasion assay), while 500 µl of medium with 10% FBS was added to the lower chamber. Then, the medium in the lower chamber was replaced with 500 µl of 10% FBS medium. After incubation for 24 h, the cells that had migrated to or invaded the bottom of the membrane were fixed with methanol/glacial acetic acid mixture for 30 min and then stained with 10% Giemsa. Then the pictures of stained cells were taken with an inverted microscope. And the cell number in three randomly fields were counted. Data were collected from repeated three independent experiments.

Cell Transfection
The human BNIP3 gene was ligated into a pGCMV/MCS/IRES/EGFP/Neo vector (GenePharma, Shanghai, China) to construct a BNIP3(+) plasmid, and the shRNA sequence against BNIP3 was ligated into a pGPU6/GFP/Neo vector. The U87 and U251 cell lines were transfected with BNIP3(+)-encoding plasmids or BNIP3 shRNA and designated BNIP3(+) cells or BNIP3(-) cells, respectively. Empty vectors were used for the negative controls (NCs). Opti-MEM and Lipofectamine 3000 reagent (Life Technologies Corporation, Carlsbad, CA, USA) were used according to the manufacturer's instructions. Transfection was performed at about 80% confluency of cells in 24-well plates with Opti-MEM R I and Lipofectamine 3,000 reagent. Stable cell lines were obtained through selection using 0.4 mg/ml geneticin (G418; Sigma-Aldrich, St Louis, MO, USA). Geneticin-resistant cells were obtained after 4 or 5 weeks. Geneticin-resistant GSCs were subsequently isolated from these cells.
The cells were transfected with a miR-24-3p agomir, miR-24-3p antagomir, or the appropriate NC (GenePharma, Shanghai, China) using Lipofectamine 3,000 reagent (Life Technologies Corporation), according to the manufacturer's instructions. After 6 h of transfection, the medium was removed and replaced with fresh medium. The cells were harvested after 48 h of transfection.

Immunofluorescence Staining
Cells were stained with LysoTracker Red or MitoTracker Red at a final concentration of 50 nM. Then, the cells were fixed in 4% paraformaldehyde for 30 min, blocked with 5% BSA for 2 h, and incubated with the primary antibody for microtubule-associated protein one light chain 3 (LC3, 1:150, Abcam, Cambridge, MA, USA, ab51520) at 4 • C overnight. Then, the cells were washed and incubated with the appropriate secondary antibody before being counterstained with DAPI (Beyotime Institute of Biotechnology, Jiangsu, China) for 10 min and visualized with a confocal microscope.

Luciferase Reporter Assays
For the luciferase reporter assays, HEK-293T cells were seeded into 96-well plates. These cells were co-transfected with a wildtype (Wt) or mutant (Mut) pmirGLO-BNIP3 reporter plasmid and an agomir-24-3p or agomir-24-3p-NC. After 48 h, luciferase activity was measured using Dual-luciferase Reporter Assay System (Promega, Madison, WI, USA). Relative firefly luciferase activity levels were calculated and normalized to that of renilla luciferase.

RNA Extraction and Quantitative RT-PCR (qRT-PCR)
Total RNA was separated from U87-GSCs and U251-GSCs using Trizol reagent (Life Technologies Corporation). A Taq-Man MicroRNA Reverse Transcription Kit was used for miRNA reverse transcription (Applied Biosystems, Foster City, CA, USA). QRT-PCR was performed using TaqMan Universal Master Mix II with Taq-Man microRNA assays for miR-24-3p and U6 expression. For quantification of BNIP3 mRNA, we performed reverse transcription and RT-PCR amplification using an RT-PCR kit and SYBR Premix Dimer Eraser (TaKaRa, Dalian, China) and the assays for BNIP3 and GAPDH gene expression (Applied Biosystems). Fold changes were calculated using the relative quantification (2 − Ct ) method.
Athymic nude mice (BALB/C-nu/nu, 4 weeks old, male) were purchased from the Cancer Institute of the Chinese Academy of Medical Science. The ethics committee of Shengjing Hospital has approved this study. The animals were raised in accordance with the guidelines of the Laboratory Animal Centre. The cells were subcutaneously implanted into the right flanks of the mice at a density of 5 × 10 5 cells/mouse (n = 10 each group). Tumor volumes were measured every 5 days until 50 days postinoculation and were calculated using the following formula: volume (mm 3 ) = length × width 2 /2.
For intracranial orthotopic inoculation, 3 × 10 5 cells were stereotactically implanted in the right striatum of the mice (n = 10 each group). Mice died within 7 days post-operation were considered to have unrelated to tumors and eliminated from the study. New mice were then treated and added to the groups. The numbers of surviving mice were recorded daily until 50 days post-operation. Survival analysis was performed using Kaplan-Meier survival curves. Tumor-bearing mice were divided into the following four groups: (1) Control group, whose mice treated with 0.9% sodium chloride; (2) EMAP-II+TMZ group, whose mice were treated with 400 µM TMZ after pretreatment with 0.05 nM EMAP-II for 0.5 h; (3) miR-24-3p(-) group, whose mice were treated with cells stably expressing miR-24-3p(-); (4) EMAP-II+TMZ+miR-24-3p(-) group, whose mice were treated with cells stably expressing miR-24-3p(-) and 400 µM TMZ after pretreatment with 0.05 nM EMAP-II for 0.5 h.

Transmission Electron Microscopy (TEM)
The cells were fixed in 2.5% glutaraldehyde (electron microscopy grade) in phosphate-buffered saline (PBS) at 4 • C for 2 h, dehydrated in an ethanol series and embedded in Epon resin. Ultrathin sections were then subjected to TEM analysis. Representative areas were chosen for ultrathin sectioning and viewed with a Philips EM 400 transmission electron microscope at an accelerating voltage of 80 kV, and digital images were obtained by an AMT imaging system (Advanced Microscopy Techniques Cor, Danvers, MA, USA).

Statistical Analysis
Data are presented as the mean ± standard deviation (SD) and were analyzed using SPSS19.0 (SPSS, IL, USA). Statistical analysis was performed using Student's t-test and one-way ANOVA. P < 0.05 was considered statistically significant. IC50 was measured by GraphPad Prism software (raw data in Data Sheet 1).

De-Differentiation and Identification of GSCs
The result we already published showed that the isolated GSCs using sphere culture by flow cytometry analysis contained 98.2 and 98.6% CD133+ cells in GSCs-U251 and GSCs-U87, respectively . To identify GSCs, we use immunofluorescence staining. The positive staining of Nestin and CD133 confirmed that most cells within the spheres expressed these neural stem cell lineage markers on their membranes (Supplementary Figure 1A). Besides, the cell spheres are differentiated into a non-GSC line in DMEM supplemented with 10% FBS, without EGF or FGF. These cells showed typical morphological differentiation toward neuronal and astrocytic lineages, identified as betatubulin-III positive neurons and GFAP-positive astrocytes (Supplementary Figure 1B).

EMAP-II Intensified the Suppressive Effects of TMZ on U87-GSCs and U251-GSCs Viability, Migration, and Invasion
The cytotoxicity of TMZ and EMAP-II was measured by Cell Counting .Based on the IC50 values of TMZ against U87-GSCs and U251-GSCs (Zhou et al., 2017), we selected the concentrations of 200 and 400 µM and the time periods of 48 or 72 h for subsequent experiments. According to the results of our preliminary study (Chen et al., 2016), 0.05 nM EMAP-II and 0.5 h were selected as the optimum pretreatment concentration and time point, respectively. Viability was apparently inhibited in the EMAP-II and TMZ groups and the EMAP-II and TMZ combination group after 48 and 72 h of treatment (P < 0.05, P < 0.05, P < 0.01). Compared with the EMAP-II and TMZ groups, the EMAP-II and 200 and 400 µM TMZ combination groups displayed significantly reduced cell viability after 48 h of treatment (P < 0.05, P < 0.01). The EMAP-II and 200 or 400 µM TMZ combination groups also displayed significantly lower cell viability than the EMAP-II group after 72 h of treatment (P < 0.05; Figure 1A). Thus, 0.05 nM EMAP-II for 0.5 h and 400 µM TMZ for 48 h were the pretreatment and treatment concentrations and times, respectively, used in subsequent experiments.
Next we explore the combination of EMAP-II and TMZ on migration and invasion of GSCs.U87-GSC and U251-GSC migration and invasion were significantly inhibited in the EMAP-II and TMZ groups and the EMAP-II and TMZ combination group (P < 0.05). Compared with the EMAP-II and TMZ groups, the EMAP-II and TMZ combination group displayed significantly decreased migration and invasion (P < 0.05; Figure 1B).

Combination Treatment With EMAP-II and TMZ Induced Mitophagy in GSCs
TEM and Immunofluorescence assay was used to monitor mitophagy. Autophagysome activity levels were increased in the EMAP-II and TMZ groups and the EMAP-II and TMZ combination group. The EMAP-II and TMZ combination group displayed more autophagysomes than the EMAP-II and TMZ groups (Figure 2A). U87-GSC and U251-GSC were stained for LC3 (green) and Mito-Tracker or Lyso-Tracker (red). The punctate density and distribution indicates LC3 expression. LC3 punctuate staining was increased in the EMAP-II and TMZ groups and the EMAP-II and TMZ combination group. LC3 punctuate staining was dramatically increased in the EMAP-II and TMZ combination group compared with the EMAP-II and TMZ groups. Additionally, we noted an overlap between LC3 and Mito-Tracker or Lyso-Tracker. The latter two agents exhibited a staining trend similar to that of LC3, and their dots colocalized with the LC3 dots (Figures 2B,C).
Western blot was used to detect the protein expression of LC3 II/I, P62, mitochondrial matrix protein CYPD and mitochondrial membrane protein TOMM20 in U87-GSCs and U251-GSCs. The LC3-II/I ratio gradually increased in the EMAP-II and TMZ groups and the EMAP-II and TMZ combination group (P < 0.01), but P62, TOMM20 and CYPD expression levels displayed gradual decreases in the EMAP-II and TMZ groups and the EMAP-II and TMZ combination group (P < 0.05). The LC3-II/I ratio increased (P < 0.05), but P62, TOMM20 and CYPD expression levels decreased in the EMAP-II and FIGURE 1 | The effects of TMZ and EMAP-II on U87-GSCs and U251-GSCs viability, migration and invasion. (A) Cytotoxic effects of the combination of EMAP-II and TMZ on U87-GSCs and U251-GSCs compared with the control treatment and EMAP-II and TMZ alone. The combination of EMAP-II and TMZ significantly inhibited cell viability of U87-GSCs and U251-GSCs compared with the control treatment and EMAP-II and TMZ alone. Data are presented as the mean ± SD from three independent experiments. (B) GSC cell migration and invasion were investigated by transwell assays after treatment with EMAP-II and TMZ. Compared with the control treatment and EMAP-II and TMZ alone, the migration and invasion of GSC are inhibited in the combination of EMAP-II and TMZ group. Data are presented as the mean ± SD of n = 3. *P < 0.05, **P < 0.01 vs. control group, # P < 0.05, ## P < 0.01 vs. TMZ group, & P < 0.05, && P < 0.01 vs. EMAP-II group. Scale bars represent 40 µm.
TMZ combination group compared with the EMAP-II and TMZ groups (P < 0.05; Figure 2D).

Combination Treatment With EMAP-II and TMZ Inhibited GSCs Malignant Behavior by Upregulating BNIP3
EMAP-II and TMZ influenced the expression of BNIP3 in GSCs. BNIP3 protein expression increased significantly in the EMAP-II and TMZ groups and the EMAP-II and TMZ combination group (P < 0.05). BNIP3 protein expression increased in the EMAP-II and TMZ combination group compared with EMAP-II and TMZ groups (P < 0.05) ( Figure 3A). Proliferation, migration, and invasion of GSCs were influenced by BNIP3 overexpression and knockdown. GSCs proliferation, migration, and invasion were inhibited in the BNIP3(+) group (P < 0.05) and were enhanced in the BNIP3(-) group (P<0.05) (Figures 3B,C).
To explore whether BNIP3 displayed an important role in the combination of EMAP-II and TMZ, GSCs were treated with EMAP-II and TMZ based on whether BNIP3 was overexpressed or knocked down. The LC3II/I ratio and BNIP3 expression was increased (P < 0.05), but P62,   -II/I ratio and P62, TOMM 20, and CYPD expression levels in U87-GSCs and U251-GSCs treated with EMAP-II and TMZ. The LC3-II/I ratio increased, but P62, TOMM 20, and CYPD expression decreased in GSCs treated with EMAP-II and TMZ. Data are presented as the mean ± SD of n = 3. *P < 0.05, **P < 0.01 vs. control group, # P < 0.05 vs. TMZ group, & P < 0.05 vs. EMAP-II group. TOMM20, and CYPD expression levels were decreased in the BNIP3(+)+EMAP-II+TMZ group compared with the EMAP-II and TMZ combination group (P < 0.05). The opposite result was observed in the BNIP3(-)+EMAP-II+TMZ group compared with the EMAP-II and TMZ combination group ( Figure 4B). GSCs proliferation, migration and invasion were increased in the BNIP3(-)+EMAP-II+TMZ group compared with EMAP-II and TMZ combination group (P < 0.05) and were inhibited in the BNIP3(+)+EMAP-II+TMZ group compared with EMAP-II and TMZ combination group (P < 0.05; Figures 4C,D).

Combination Treatment With EMAP-II and TMZ Induced Mitophagy and Inhibited GSC Malignant Behavior by Down-Regulating miR-24-3p
The expression level of miR-24-3p in GSCs was detected by real-time PCR. MiR-24-3p expression levels were decreased in the EMAP-II and TMZ groups and the EMAP-II and TMZ combination group (P < 0.05). MiR-24-3p expression levels were decreased in the EMAP-II and TMZ combination group compared with the EMAP-II and TMZ groups (P < 0.05; Figure 5A). Mitophagy-related proteins were affected by miR-24-3p. The LC3-II/I ratio and BNIP3 protein expression were decreased (P < 0.05), but P62, TOMM20, and CYPD expression levels were increased in the Agomir-24-3p group compared with the corresponding Control group (P < 0.05). Conversion of LC3-I to LC3-II and BNIP3 protein expression were increased (P < 0.05), but P62, TOMM20, and CYPD expression levels were decreased in the Antagomir-24-3p group compared with the corresponding Control group (P < 0.05) (Figures 5B,C). GSC proliferation, migration and invasion were inhibited in the Antagomir-24-3p group compared with the corresponding Control group (P < 0.05) and were enhanced in the Agomir-24-3p group compared with the corresponding Control group (P < 0.05; Figures 5D,E). (Continued) FIGURE 4 | EMAP-II+TMZ group, # P < 0.05 vs. EMAP-II+TMZ group.(C,D) CCK8 and transwell assays of the effects of BNIP3 overexpression or silencing and the combination of EMAP-II and TMZ on the proliferation, migration, and invasion of U87-GSCs and U251-GSCs. The combination of BNIP3 overexpression, EMAP-II and TMZ inhibited GSCs proliferation, migration, and invasion. Data are presented as the mean ± SD of n = 3. *P < 0.05 vs. EMAP-II+TMZ group, # P < 0.05, **P < 0.01 vs. EMAP-II+TMZ group. Scale bars represent 40 µm.
To furtherly explore whether miR-24-3p could affect BNIP3 in the combination of EMAP-II and TMZ, GSCs were treated with EMAP-II and TMZ on the basis of over-express or downregulate miR-24-3p. Compared with EMAP-II+TMZ group, the expression of BNIP3 was reduced in agomir-24-3p+ EMAP-II+TMZ group; the expression of BNIP3 was increased in antagomir-24-3p+EMAP-II+TMZ group (Figure 6C).

The Combination of miR-24-3p, EMAP-II and TMZ Inhibited Tumor Growth in Vivo by Inducing BNIP3-Mediated Mitophagy
To prove the above findings, the growth-inhibitory effect of EMAP-II, TMZ and miR-24-3p inhibitor on U87-GSCs and U251-GSCs was tested in xenografted mice. The tumor volumes of the xenografts were significantly smaller in the EMAP-II+TMZ, miR-24-3p(-) and miR-24-3p(-)+EMAP-II+TMZ groups than in the Control group. The tumor volume of EMAP-II+TMZ group is smaller than that of the EMAP-II and the TMZ group. The minimum volume was noted in the miR-24-3p(-)+EMAP-II+TMZ group (Figures 8A-C). Mouse survival time were much longer in the EMAP-II+TMZ, miR-24-3p(-) and miR-24-3p(-)+EMAP-II+TMZ groups than in the Control group. Mouse survival time was significantly longer in the miR-24-3p(-)+EMAP-II+TMZ group than in the other groups ( Figure 8D).

DISCUSSION
We found that EMAP-II enhanced the cytotoxic effects of TMZ on GSC viability, migration and invasion. The combination of TMZ and EMAP-II induced GSC mitophagy, which was shown to inhibit GSC malignant behavior by upregulating BNIP3mediated mitophagy. We also found that the combinaltion of TMZ and EMAP-II decreased miR-24-3p expression and that BNIP3 was the target gene of miR-24-3p. The combination of TMZ and EMAP-II induced mitophagy to inhibit malignant GSC behavior by down-regulating miR-24-3p. The results of our experiments with the xenografted mice showed that the EMAP-II+TMZ, miR-24-3p(-) and miR-24-3p(-)+EMAP-II+TMZ groups had smaller tumors and longer survival times than the other groups. In particular, the miR-24-3p(−)+EMAP-II+TMZ group had the smallest tumors and the longest survival time.
FIGURE 6 | BNIP3 was the direct target of miR-24-3p, and its expression level changed after miR-24-3p overexpression or silencing in U87-GSCs and U251-GSCs. (A) Relative luciferase activity was expressed as firefly/renilla luciferase activity. Luciferase activity was significantly decreased in the Wt BNIP3 3 ′ untranslated region (3 ′ UTR)+agomir-24-3p group. Data are presented as the mean ± SD of n = 3. *P < 0.05 vs. agomir-24-3p-NC group. (B) BNIP3 relative mRNA expression levels were detected by qRT-PCR in U87-GSCs and U251-GSCs after miR-24-3p overexpression or silencing. MiR-24-3p overexpression or silencing do not influence BNIP3 mRNA expression. Data are presented as the mean ± SD of n = 3. (C) Western blot analysis of BNIP3 levels in U87-GSCs and U251-GSCs treated with EMAP-II and TMZ on the basis of miR-24-3p overexpression or silencing. Compared with EMAP-II+TMZ group, the expression of BNIP3 was reduced in agomir-24-3p+ EMAP-II+TMZ group. The opposite result was in antagomir-24-3p+EMAP-II+TMZ group. Data are presented as the mean ± SD of n = 3, **P < 0.01 vs. EMAP-II+TMZ group, *P < 0.05 vs. EMAP-II+TMZgroup. was affected by TMZ in a time-dependent manner over 24, 48, and 72 h of treatment. In addition, we found that the suppressive effects of TMZ on GSCs were also dose-dependent. However, when the TMZ concentration was 400 µM or higher, this dose-dependent effect was no longer observed, a finding consistent with those of other studies. Yu et al. found that the cytotoxic effects of TMZ on U251-GSCs and U87-GSCs were time-and dose-dependent and that 200 µM TMZ treatment for 48 h could decrease cell viability by almost 30% (Yu et al., 2015). In recent years, studies have shown that EMAP-II plays an anti-tumor role through several processes, as the agent inhibits angiogenesis, promotes endothelial cell apoptosis and directly promotes tumor cell autophagy and apoptosis (Crippa et al., 2008;Li Z. et al., 2014;Ma et al., 2015;Chen et al., 2016). EMAP-II is still in the experimental study stage at present. Some studies have found that EMAP-II could facilitate TNF-R1 apoptotic signaling, which exerts anti-tumor functions (van Horssen et al., 2006). It also increases the permeability of blood-tumor barrier. Both are of benefit to killing tumor cells. Therefore, we think that EMAP-II has a good applied future in the clinical practice. However, whether EMAP-II can enhance the cytotoxicity of TMZ remains unclear. It has been reported that low-dose EMAP-II generated cytotoxic effects through the PI3K/Akt/FoxO1 axis in U87-GSCs . TMZ could activate the Akt/glycogen synthase kinase-3ß pathway to induce glioma cell apoptosis via oncoprotein c-Myc (De Salvo et al., 2011) and exerted anti-tumor effects through the AKT-mTOR pathway in the GBM8901 glioma cell line . EMAP-II and TMZ affected glioma cells through a similar pathway, indicating that EMAP-II may strengthen the cytotoxic effects of TMZ. Our team have proved that the combination of EMAP-II and TMZ kill GSCs by inducing autophagy. Autophagy inhibitors weaken the cytotoxic effect of EMAP-II and TMZ on GSCs. Our research results are consistent with previous studies (Zhou et al., 2017). We also found that EMAP-II and TMZ kill GSCs through inducing mitophagy. We found that the cytotoxic effects of TMZ on GSCs were significantly enhanced and that GSC migration and invasion were also inhibited after EMAP-II pretreatment, but the concrete mechanisms underlying these effects remain to be investigated further.
Mitophagy plays an important role in a variety of diseases, such as diabetes, cancer and degenerative diseases of the nervous system (Greene et al., 2003;Soleimanpour et al., 2015;Fader et al., 2016;Wilfinger et al., 2016). Consequently, mitophagy has been regarded as a target in anti-tumor therapy. Recently, some chemotherapy drugs were also found to exert anti-tumor effects by inducing mitophagy. For example, salinomycin promoted mitophagy to inhibit the growth of breast cancer and prostate cancer cells (Jangamreddy et al., 2013), and sodium selenite induced superoxide-mediated mitophagy and subsequent autophagic cell death in malignant glioma cells (Kim et al., 2007). EMAP-II can induce mitophagy in the U87, U118 and GSCs cell lines (Ma et al., 2015), but the specific mechanism underlying this effect remains unclear. TMZ has been found to induce mitochondrial-mediated autophagy and reduce the quantity of mitochondria in the U87 cell line (Lin et al., 2012). Our results showed the combination of EMAP-II and TMZ promoted the formation of autophagosome by transmission electron microscopy and increased the overlap between LC3 and Mito-Tracker or Lyso-Tracker by immunofluorescence staining. EMAP-II and TMZ up-regulated LC3II/I ratio and BNIP3 expression but down-regulated P62, TOMM20, and CYPD expression, indicating promoting mitophagy. These results indicated that the combination of EMAP-II and TMZ inhibited GSC malignant behavior by enhancing mitophagy.
BNIP3 is distributed in the mitochondrial outer membrane and interacts with LC3 through the LC3-interacting region (LIR) to trigger excessive mitophagy (Hanna et al., 2012;Zhu et al., 2013). Arsenic trioxide and ceramide can upregulate BNIP3 expression to induce mitophagy in glioma (Daido et al., 2004;Kanzawa et al., 2005). We confirmed that the combination of EMAP-II and TMZ inhibited the biological behavior of GSCs by upregulated BNIP3 expression. BNIP3 overexpression enhanced the effect of EMAP-II and TMZ, which up-regulated the LC3II and down-regulated the expression level of P62, TOMM20 and CYPD, and inhibited the biological behavior of GSCs. BNIP3 knockdown exerted the opposite effect.
The results of the above demonstrated the combination of EMAP-II and TMZ induced mitophagy by upregulating BNIP3 expression. BNIP3 overexpression could directly enhance the synergistic anti-tumor effects of EMAP-II and TMZ by inducing mitophagy. However, BNIP3 knockdown exerted the opposite effects. These results indicated that BNIP3-mediated mitophagy is one of the anti-tumor mechanisms underlying the effects of the combination of EMAP-II and TMZ. Abnormal increases in BNIP3 expression are affected by HIF-1 expression, which is positively correlated with tumor metastasis in non-small cell lung cancer (Giatromanolaki et al., 2004). BNIP3 silencing caused by promoter methylation and histone acetylation is associated with chemotherapy resistance and worse survival, and BNIP3 restoration enhances chemosensitivity and promotes tumor cell apoptosis in gastric and colon cancer (Murai et al., 2005). These results indicates BNIP3 loss caused by various reasons is related with tumor progression, which consists with our results.
MiRNAs plays key roles in tumourigenesis and tumor progression as oncomirs (oncogenic miRNAs) or tumor suppressors. EMAP-II upregulated ATG7 and ATG5 to enhance U87 and U251 glioma cell autophagy via miR20a downregulation (Chen et al., 2016). It has also been reported that TMZ upregulated ATG7 expression to induce autophagy by down-regulating miR-17 (Comincini et al., 2013). Previous study show miR-24-3p promoted cell proliferation and inhibited apoptosis in human breast cancer cells by targeting p27Kip1 (Lu et al., 2015). Moreover, miR-24-3p promoted cell proliferation in glioma cells via cooperative regulation of MXI1 (Xu et al., 2013). The results of these studies suggested that miR-24-3p may serve as an oncomir. And can EMAP-II and TMZ regulated mitophagy via miR-24-3p? Our results showed that the combination of EMAP-II and TMZ reduced miR-24-3p expression. MiR-24-3p overexpression decreased the LC3-II/I ratio and increased P62, TOMM20, and CYPD expression levels. MiR-24-3p overexpression also promoted the abilities of GSC proliferation, migration, and invasion. While miR-24-3p knockdown exerts the opposite effects. All the results indicated that combination treatment with EMAP-II and TMZ induced mitophagy and inhibited GSC malignant behavior by downregulating miR-24-3p.
MiRNAs can bind to the 3 ′ -UTRs of target mRNAs and block their translation. MiR-137 inhibited mitophagy by regulating the mitophagy receptors FUNDC1 and NIX (Li W. et al., 2014). BNIP3 was found to be the target gene of miR-24-3p using bioinformatics software. Dual-luciferase reporter assay illustrated that miR-24-3p binds with the 3 ′ -UTR of BNIP3 mRNA. BNIP3 overexpression reversed the effects of agomir-24-3p on the LC3 II/I ratio and P62, TOMM 20, and CYPD expression, and BNIP3 knockdown reversed the effects of antagomir-24-3p on the LC3 II/I ratio and P62, TOMM 20, and CYPD expression. These resultsillustrate that miR-24-3p negatively regulates BNIP3, which inhibits mitophagy and promotes GSC proliferation, migration and invasion. Furthermore, we also found that combination of miR-24-3p(-), EMAP-II and TMZ inhibited tumor growth in vivo.
In summary, we have provided strong experimental evidence that EMAP-II can enhance the inhibitory effects of TMZ on GSC viability, migration and invasion. Regarding the mechanisms underlying this effect, the combination of TMZ and EMAP-II upregulated BNIP3 expression by down-regulating miR-24-3p, thereby strengthening BNIP3-mediated mitophagy. Because of its anti-tumor effects, the combination of EMAP-II and TMZ may represent a novel therapeutic strategy for treating glioma.

ETHICS STATEMENT
The animals were raised in accordance with the guidelines of the Laboratory Animal Centre of Shengjing Hospital.